Schizophrenia is a common and severe psychiatric disorder whose etiology appears to be substantially related to genetic factors. Genetic linkage studies of families with multiple cases of schizophrenia-related disorders have begun to produce some convergence of evidence for the location of susceptibility genes in specific chromosomal candidate regions. However, evidence is inconsistent across studies, genome-wide statistical significance has been obtained in only two regions in single samples, and no susceptibility locus has been identified. One possible explanation is that each susceptibility locus has either a very small effect on risk in most families or a moderate effect on risk in a small proportion of families; in either case, the measured effect across a population will be small. One strategy for improving the detection of such loci is to study very large samples of multiply-affected families, which is currently possible only through multicenter collaborations. Seven investigators with a record of successful collaboration propose to undertake three studies over the next three years to investigate schizophrenia candidate regions in eight samples (their own plus the NIMH Schizophrenia Genetics Initiative sample which is publicly available). The combined sample includes 866 pedigrees containing 2,168 individuals affected with schizophrenia or schizoaffective disorder (DSM-IIIR), with 901 independent affected sibling pairs (ASPs) (with s-1 correction). During the first year, candidate regions on chromosomes 15q, 18p and 22q will be studied by genotyping maps of microsatellite markers at approximately 5 cM spacing across each candidate region (an estimated 33 markers), with appropriate quality control procedures. Multipoint ASP and NPL analyses will be performed, in addition to a logistic regression method to take site differences into account. Concurrently, a new rank order meta-analysis method will be applied to all available schizophrenia genome scan data to identify additional candidate regions. In years 2 and 3, two additional studies of a similar design will be undertaken in candidate regions identified either by the meta-analysis or on the basis of new reports in the field. These studies may assist investigators in determining which candidate regions are most deserving of intensive efforts to find disease-related genes, and the most likely locations of these genes within the regions.